. 2021 Oct 30;10:285. doi: 10.1186/s13643-021-01763-w

Table 3.

Performance of srBERT_my with respect to the learning steps for the second task (relation extraction) using datasetB

	srBERT_my50K	srBERT_my100K	srBERT_my150K	srBERT_my200K	srBERT_my250K	srBERT_my300K	srBERT_my350K	srBERT_mix	Original BERT
Accuracy	0.922	0.935	0.896	0.909	0.922	0.909	0.909	0.922	0.922
Loss	0.337	0.270	0.542	0.540	0.328	0.658	0.658	0.309	0.232

SR systematic review, BERT bidirectional encoder representations from transformers, srBERT_my#K srBERT_my model trained for # K steps